5 Conclusions

After this review-like study, a result of reading more than 200 publications, reports, proceedings and other papers, the following conclusions can be distilled.

It is generally agreed that half of the world’s population lives in a landscape that can be classified as ‘urban’. However, this does not mean that half of the world lives in cities. In large parts of the world, especially in developed countries or rapidly developing countries, these urban landscapes differ in density, spatial configuration and land use properties from ‘urban cores’, cities or towns. The increasing number of papers on the broad theme of intake of rural and natural area by urbanization processes and of the resulting state of a semi-urbanized conditions and its impacts on society, the economy, and the environment, leaves no doubt that this subject is soliciting due interest. The major conclusion of this review, however, is the apparent persistent lack of consistency in definitions, vocabulary, and above all, a diagnostic grammar.

The two most frequently used concepts to describe the semi-urban state are “fringe” and “sprawl”. Although the semantics are somewhat blurred, both terms point to fundamentally different spatial expressions of urban expansion. Fringe implies an outer zone of urban expansion that outwardly radiates from core urban areas. Sprawl much more implies the emergence of spatially distributed urban conditions, much less dependent of core areas.

Research on sprawl, fringes, and semi-urban areas can is conducted in ecological, social as well as economical sciences; it leaves no doubt that the whole sprawl-fringe-semi-urban area concourse is a full-issue sustainability topic. In this paper, only the ecological dimension of the sustainable development is the primary focus. When browsing literature about sprawl, urban fringes, and semi-urban landscape, ecological impacts are most commonly studied, since these forms of land cover dynamics can have great impacts on biomass availability, biodiversity, soil quality, hydrology, landscape structure and conservation, habitat fragmentation and many other topics. A second topic, often found linked with sprawl and semi-urban areas are technical issues, frequently dealing with data properties like scale effects. The need for high resolution and very detailed data to study semi-urban areas and sprawl effects is repeatedly emphasized.

A broad range of development strategies and concepts exists, almost all of them have as goal sustainable (urban) planning and propose, for example, measures to restrict sprawl: New Urbanism and Landscape Urbanism as architectural theories, Smart Growth as an example of sustainable policy concept.

One would expect an equilibrated approach to semi-urban areas from two groups of disciplines: the “urban disciplines” and the “rural disciplines”. To the contrary, however, there is a very strong bias of urbanistic approaches. This can be explained by the fact that the most obvious and active process in these areas is of strong urban nature, the rural and natural area being the “loosing aspect”. Whereas the phenomena of sprawl and the fringe can be fully acknowledged as specific forms of urban development, the residual rural characteristics in the semi-urban area seem to be much less attractive as focal issues for rural development. So, from international to local level, the definition of rurality commences with the exclusion of land areas above a threshold of urban development. Urbanistic theories and practice are much less tempted to exclude thinly built-up areas from their fields of research and application.

Exceptions to this exclusion behavior of rural research and application can be found in the upcoming theme of urban agriculture, but here again, the key characteristic of this agriculture is urban, and the emphasis in this field is to turn farming as part and parcel of urban, not rural development. This apparently persistent allocating of open space functions to either the rural or the urban side of the semi-urban continuum appears to be difficult to be countered. The concept of neo-rurality, although the name suggests positioning at the rural side, is meant to develop concepts for sustainable use of unbuilt spaces irrespective of their geographical position, close to or far away from urban fabric.

Much less frequently, compared to sprawl and fringe, half urbanized areas are being categorized as a specific type of landscape. This can be explained in different ways. First of all, the term of landscape implies some recognizable order or pattern in terms of ecological functioning, land use, visual quality, historic development or characteristics in general sense. The typical heterogeneity of spatial, environmental and use conditions in half urbanized areas make these hard to fit into the landscape analytical schemes. However, the landscape concept – with its branches in landscape ecology, landscape design, etc. – is a potential field for developing sustainability development principles, but here again the urban–rural divide is lurking, for instance in the discipline of landscape urbanism, in which landscape principles are fully integrated in urban development.

Although plenty of research is committed to topics such as sprawl, urban fringes, and semi-urban areas, several authors point to the lack of a clear definition of what sprawl, the urban fringe, and a semi-urban area is.

This paper illustrates that sprawl, urban fringe dynamics, and semi-urban area issues are ‘hot topics’ in actual research concerning sustainable development. It also illustrates the great diversity of concepts, theories, definitions (and the lack of them) and approaches. For an important part, this can be justified by the fact that every scientific discipline starts from its own basis and thus develops its own approach and theory around the topics here discussed. On the other hand, this is in contradiction with actual popular concepts of integrated, inter- and multidisciplinary approaches, certainly in landscape research (Antrop, 2000a; Pickett et al., 2001; Tress et al., 2006; Wu and Hobbs, 2002). Where there is already a fair amount of research done on the impacts of urbanizing and already (semi-) urbanized landscapes and the application of existing planning concepts as remedies, the construction of a general, multi-applicable framework for landscapes subjected to urban sprawl, the urban fringe, or semi-urban landscapes is still scarcely out of the egg.

Towards a general definiton for semi-urban areas

An important next step in the aftermath of this study, could be the formulation of a new, generalizing definition of semi-urban areas, although one of the main conclusions of this paper is that finding a “general denition” is very difficult, maybe even impossible, due to the various different approaches on the subject, the different frameworks, needs, etcetera. As an illustration, a possible definition – or at least start of – is provided in the next paragraphs. It has to be noted that this should be understood as a beginning and that the definition is not (yet) tested on a sufficient number of case study areas and is highly dependent on the (landscape ecological) background of this study.

In a landscape ecological context, a landscape can be identified as a semi-urban area,

• when its land cover consists of a sufficient high degree, but not exclusively, of artificial land use with a mainly urban land cover that can be classified as sealed surface (transport infrastructure, buildings and other built-up urban fabric) and
• when its land cover also consists of a sufficient degree, but not exclusively, of green structures, possibly entirely or partly of artificial nature, that may accompany built-up structures or have a specific social or ecological function in an urban context, like parks, road verges, gardens and
• when it can be clearly measured that the land use is in some way dynamic, either because of a pressure from factors coming from other surrounding landscapes (population growth or movement, for example) or a pressure created by the land use itself (for example, more intense road infrastructure development because of a higher degree of commuting) and
• when it can also be proven, possibly by using a reference landscape, that the area of study is neither “pure” urban core (village, city, town), nor “pure” agricultural” land, nor “pure” natural area. In principle, if the triangular relationship between (a), (b) and (c) is clearly a fact, condition (d) is unnecessary.